Chat GPT

[Andy_W]

What automations are you trying to do that covers most of your work bud?

To do it I'm basically going to make a program in Python (seems really easy to learn the basics of what I need) which I can add modules too, using a basic GUI like TKinter. I've just been using chat GPT to help with some of the coding really as well as youtube etc. Started it a couple of months back, and made some good progress, but ran out of time, so lost my way with it a bit, but going to get back on with it soon. I think I know how to do everything I want now (or I did), I just need the time to do it really.

All of this goes into a database, so I can call up various jobs and options within that.

I've done all the aspects of it separately in various ways over the years using Excel, but it's slow, repetitive and not particularly user-friendly. I'm the only user mind, but intending to sell up soon, so will need to be able to do things quickly whilst I'm sorting out a handover and passing over what I've done now for someone else to use would basically be impossible.

I'll possibly end up using the same thing going forward as a consultant, but won't likely be handing over the program I create

 

[Andy_W]

It is really like talking to a psuedo expert on everything Andy, are you sue you are not a chatbot?

I am not arguing against using it for my use. I am arguing against using it for any coding that has any impact. It is often hopeless and can produce code that won't run, or worse still run, but not do what you want it to. It is hopeless. It can't even organize a project for VS creating stub classes. It's hopeless mate.

As for beginer coders, not sure about where your mates work, but we don't let a graduate anywhere near live code for about a year after graduating. Thats precisely because we write complex code. So no, I am not proving your point.

It's mad how you think that only your world exists, good luck with it

 

[Thing]

All good fun and games but ask yourselves where this is going. Apologies for urinating on a previously enligthening post. I will get lambasted for this but hey ho. Helmet on. Fire away. Thick skinned.

At the minute AI can write some noddy code, predict the scores for the weekend, make up a story, help with the kids homework..... all harmless fun. BUT.........

The market is totally unregulated so there are no restrictions on what it is developed for, what it can learn, what it can do with the data it is gleaning. I am torn at the minute as I have been asked to lead on a project to implement AI into a workspace that would ultimately see people potentially be removed from their roles. And this is where it starts getting scary.

Three of the pillars of our civilisation are education, law and order, and democracy. Agreed?

I've already had to remove the iPad from one of my kids when I found him using chatGPT to do his homework - we're not talking about helping him, it wrote the entire essay that he was then copying word for word. Some will say that is harsh, some will say I'm too draconian but how does he become educated if his iPad does it for him? I don't want my kids growing up totally reliant on technology to think, formulate ideas, and express themselves. This next generation is already totally phuqqed up enough with being molly-coddled, pampered, and endowed with an overriding sense of self entitlement as it is.

There are a significant amount of white collar jobs (and that is an important point) that could be replaced by AI bots. Not the traditional blue collar jobs that have been often impacted by the advancement of technology, but well paid white collar workers. We will see entire professions decimated over the next quarter century - high paid organisational resource (lawyers and accountants) being prime candidates - thse who deal with facts and rules within the bounds of which code can interpret. Millions of people will no longer be able to support themselves or their families. A signifcant source of tax income will have been lost, and the double-whammy the state will then have to support those individuals who are no longer paying the tax. This will affect services as money dries up, standard of living drops, disgruntlement and depression sets in, civil unrest soon follows..... the rest has been repeated over and over again since time began.

What data is your AI bot of choice collecting about you? Combined with what social media, Amazon, et al, already have then that IS being used to manipulate what you see, what you think, what you do. We take for granted that adverts are based on our vieweing and purchase history, but when that data is used to influence how you vote then democracy is being weakened by those with the power to use that data. My next rage against the machine wil be around geo-political hacking and the fact that from now on no democratic vote can ever be trusted....but that is for another day.

Pandora's box has been opened and the lid ain't going back on!

 

[Laughing]

It's mad how you think that only your world exists, good luck with it

Now Andy its mad that you think your view is valid when talking to someone with a life times experience in this field.

I have done several open ai implementations but you, as always, think you're views are somehow equal they are not.

 

[Laughing]

All good fun and games but ask yourselves where this is going. Apologies for urinating on a previously enligthening post. I will get lambasted for this but hey ho. Helmet on. Fire away. Thick skinned.

At the minute AI can write some noddy code, predict the scores for the weekend, make up a story, help with the kids homework..... all harmless fun. BUT.........

The market is totally unregulated so there are no restrictions on what it is developed for, what it can learn, what it can do with the data it is gleaning. I am torn at the minute as I have been asked to lead on a project to implement AI into a workspace that would ultimately see people potentially be removed from their roles. And this is where it starts getting scary.

Three of the pillars of our civilisation are education, law and order, and democracy. Agreed?

I've already had to remove the iPad from one of my kids when I found him using chatGPT to do his homework - we're not talking about helping him, it wrote the entire essay that he was then copying word for word. Some will say that is harsh, some will say I'm too draconian but how does he become educated if his iPad does it for him? I don't want my kids growing up totally reliant on technology to think, formulate ideas, and express themselves. This next generation is already totally phuqqed up enough with being molly-coddled, pampered, and endowed with an overriding sense of self entitlement as it is.

There are a significant amount of white collar jobs (and that is an important point) that could be replaced by AI bots. Not the traditional blue collar jobs that have been often impacted by the advancement of technology, but well paid white collar workers. We will see entire professions decimated over the next quarter century - high paid organisational resource (lawyers and accountants) being prime candidates - thse who deal with facts and rules within the bounds of which code can interpret. Millions of people will no longer be able to support themselves or their families. A signifcant source of tax income will have been lost, and the double-whammy the state will then have to support those individuals who are no longer paying the tax. This will affect services as money dries up, standard of living drops, disgruntlement and depression sets in, civil unrest soon follows..... the rest has been repeated over and over again since time began.

What data is your AI bot of choice collecting about you? Combined with what social media, Amazon, et al, already have then that IS being used to manipulate what you see, what you think, what you do. We take for granted that adverts are based on our vieweing and purchase history, but when that data is used to influence how you vote then democracy is being weakened by those with the power to use that data. My next rage against the machine wil be around geo-political hacking and the fact that from now on no democratic vote can ever be trusted....but that is for another day.

Pandora's box has been opened and the lid ain't going back on!

You are right the industry is unregulated, but let me try and address some of your concerns.


You are right, your kids, or anyone elses shouldn't be cutting and pasting from ChatGPT. There are tools available to check for plagiarism at uni, they will be expanded to schools, moving forward. To counter this, a few years ago there was the same concerns about wikipedia (though not encyclpedia britanica, oddly enough). A teacher set some homework. His students had to discuss the history of the channel tunnel. He edited the wiki page to say that the tunnel was started in napoleonic times and was only wide enough to get a horse and cart through, and was later widened to accomodate trains. He wanted to see how many of his students did a wiki hatchet job. The answer was, most of em.

AI use in education may be a good thing, maybe not, it depends how it's harnassed. It's use at Uni will be very limited where the work is graded in a very specific way. Can you understand teh material, can you apply the material in one set of circumstances and finally can you generalise the material to new situations. AI won't be able to do #3 anytime soon.

In the workplace the industrial revolution threatened blue collar workers, the AI revolution will threaten white collar workers. Not a lot we can do about this. In the short term not much will change. Over time we won't need accountants, programmers, barristers, judges, doctors. Parliament will become a quaint footnote in our history with decentralised decision making by us all.

One final thought, and I did mention this earlier. We are a ,country mile away from general AI. It is an incredibly difficult problem to be solved because we don't understand how it works. Some biologists think consciousness arises spontaneously in any complex system. Scary thought and not everyone agrees with that, it's a guess. However, if the problem is solved, given humans aren't evolving anymore and computers are, it wouldn't take too long for AI to surpass human intelligence. This will not happen in my lifetime. Whilst ChatGPT seems impressive, it really isn't, at it's core. It's a natural language google.

 

[Thing]

You are right the industry is unregulated, but let me try and address some of your concerns.


You are right, your kids, or anyone elses shouldn't be cutting and pasting from ChatGPT. There are tools available to check for plagiarism at uni, they will be expanded to schools, moving forward. To counter this, a few years ago there was the same concerns about wikipedia (though not encyclpedia britanica, oddly enough). A teacher set some homework. His students had to discuss the history of the channel tunnel. He edited the wiki page to say that the tunnel was started in napoleonic times and was only wide enough to get a horse and cart through, and was later widened to accomodate trains. He wanted to see how many of his students did a wiki hatchet job. The answer was, most of em.

AI use in education may be a good thing, maybe not, it depends how it's harnassed. It's use at Uni will be very limited where the work is graded in a very specific way. Can you understand teh material, can you apply the material in one set of circumstances and finally can you generalise the material to new situations. AI won't be able to do #3 anytime soon.

In the workplace the industrial revolution threatened blue collar workers, the AI revolution will threaten white collar workers. Not a lot we can do about this. In the short term not much will change. Over time we won't need accountants, programmers, barristers, judges, doctors. Parliament will become a quaint footnote in our history with decentralised decision making by us all.

One final thought, and I did mention this earlier. We are a ,country mile away from general AI. It is an incredibly difficult problem to be solved because we don't understand how it works. Some biologists think consciousness arises spontaneously in any complex system. Scary thought and not everyone agrees with that, it's a guess. However, if the problem is solved, given humans aren't evolving anymore and computers are, it wouldn't take too long for AI to surpass human intelligence. This will not happen in my lifetime. Whilst ChatGPT seems impressive, it really isn't, at it's core. It's a natural language google.

An unexpected, knowledgable and articulate response. I thank you good sir.

 

[SmallTown]

An unexpected, knowledgable and articulate response. I thank you good sir.

Seconded, thanks @Laughing that was a really good breakdown on a subject most of us (certainly me) are ignorant of and potentially worried by.

 

[Littlejimmy]

I'm not worried that much about it having spent the morning wondering why my Outlook spontaneously and inexplicably changed my folder view. Computers are still full of bugs and need humans to help them out with a hard reset once in a while. From what I've seen it's just a fancy google/Siri/Alexa.

 

[Thing]

"rewrite snow white and the seven dwarves with snow white as an axe murderer"

Once upon a time, in a dark and twisted forest, there lived a wickedly captivating woman named Snow White. She possessed a chilling beauty that lured unsuspecting victims into her web of terror. Instead of being the innocent princess we know, this Snow White had a secret—a penchant for wielding an axe with deadly precision.
Snow White's stepmother, the evil queen, was consumed by envy and despised her stepdaughter's allure. In her twisted mind, she concocted a wicked plan to eliminate Snow White once and for all. But little did the queen know that Snow White had a darkness within her that matched her own.
As the tale unfolds, the queen sent the huntsman to slay Snow White in the heart of the forest. But much to the huntsman's surprise, he found himself face-to-face with a cunning and ruthless Snow White, clutching an axe stained with the blood of her previous victims.
Instead of succumbing to the huntsman's blade, Snow White turned the tables, revealing her true nature. With a sinister smile, she struck down the huntsman, silencing his cries forever.
Snow White knew that her stepmother would never stop hunting her. In order to ensure her survival, she sought refuge deep within the forest, stumbling upon a hidden cottage that belonged to a peculiar group of seven dwarves. Unbeknownst to them, they were about to encounter a force far more dangerous than they could ever imagine.
The dwarves, initially skeptical of Snow White's intentions, soon fell under her enchanting spell. One by one, she manipulated them, using their naivety to her advantage. The cottage walls witnessed the gruesome acts committed by Snow White, as she eliminated anyone who dared to stand in her way.
Yet, despite her sinister nature, the dwarves couldn't help but feel captivated by Snow White's irresistible allure. The lines between fear and fascination blurred, as they became accomplices to her dark deeds.
News of the axe-wielding enchantress spread throughout the kingdom, and even the queen trembled at the thought of facing Snow White's wrath. Consumed by paranoia, the queen disguised herself and ventured into the forest, intending to put an end to Snow White's reign of terror.
However, the queen underestimated the depths of Snow White's malevolence. In a chilling twist, Snow White cunningly lured the queen into a trap, using her beauty to mask the horrors that awaited her. With a swift swing of her axe, Snow White struck the final blow, extinguishing the queen's wicked existence forever.
As the forest fell into an eerie silence, Snow White reveled in her newfound dominance. The dwarves, now fully under her control, served as her loyal subjects. Together, they ruled the enchanted forest with an iron fist, spreading fear and darkness to all who dared to cross their path.
And so, this dark and twisted version of Snow White became a cautionary tale—a chilling reminder that evil can manifest in the most enchanting forms. The legend of Snow White, the axe murderer, continues to haunt the hearts and minds of those who dare to hear her name.

 

[Littlejimmy]

What spooked me more was the T1000 appearing when I asked an AI drawing app to show Boro winning the league...

 

[Andy_W]

Now Andy its mad that you think your view is valid when talking to someone with a life times experience in this field.

I have done several open ai implementations but you, as always, think you're views are somehow equal they are not.

Your use is not the only use, how hard is that to understand? Relatively basic code can save companies millions of pounds or thousands of wasted man-hours.

Maybe you can tell the guys at ChatGPT, Codex, Copilot or whatever that it's all worthless/ pointless?

Of course they won't agree, and there is no doubt it's getting better every year.

 

[Nano]

I think you are taking about different things. @Laughing is correct that ChatGPT doesn't write code. It returns search results depending on what you ask it. Google does that as well but in the form of a website/forum which allows you to read it, analyse it and determine whether the info it has given you is trustworthy, reliable and/or correct. ChatGPT just gives you it as an answer and has no understanding of it. That means you can ask it something very simple and it will return an answer, which might be correct, but without expert knowledge then how can you trust it?

If you are an expert then that simple code can probably be written as quickly as it would take for ChatGPT to give you it and for you to error check. If you aren't an expert you can't trust it. It's more likely to cost you than save you money if you aren't an expert.

So yes, it can give you some basic code but no, it can't write code.

 

[1finny]

You are right the industry is unregulated, but let me try and address some of your concerns.


You are right, your kids, or anyone elses shouldn't be cutting and pasting from ChatGPT. There are tools available to check for plagiarism at uni, they will be expanded to schools, moving forward. To counter this, a few years ago there was the same concerns about wikipedia (though not encyclpedia britanica, oddly enough). A teacher set some homework. His students had to discuss the history of the channel tunnel. He edited the wiki page to say that the tunnel was started in napoleonic times and was only wide enough to get a horse and cart through, and was later widened to accomodate trains. He wanted to see how many of his students did a wiki hatchet job. The answer was, most of em.

AI use in education may be a good thing, maybe not, it depends how it's harnassed. It's use at Uni will be very limited where the work is graded in a very specific way. Can you understand teh material, can you apply the material in one set of circumstances and finally can you generalise the material to new situations. AI won't be able to do #3 anytime soon.

In the workplace the industrial revolution threatened blue collar workers, the AI revolution will threaten white collar workers. Not a lot we can do about this. In the short term not much will change. Over time we won't need accountants, programmers, barristers, judges, doctors. Parliament will become a quaint footnote in our history with decentralised decision making by us all.

One final thought, and I did mention this earlier. We are a ,country mile away from general AI. It is an incredibly difficult problem to be solved because we don't understand how it works. Some biologists think consciousness arises spontaneously in any complex system. Scary thought and not everyone agrees with that, it's a guess. However, if the problem is solved, given humans aren't evolving anymore and computers are, it wouldn't take too long for AI to surpass human intelligence. This will not happen in my lifetime. Whilst ChatGPT seems impressive, it really isn't, at it's core. It's a natural language google.

A brilliant 30 mins of this on R4 this morning (11.00) The guy (Stuart Russell - computer scientist) advises the EU and US government on AI.
His view - the world is at last waking up to a ticking time bomb and he is not optimistic they will to be sensible and regulate will prevail. The economic advantages of AI are likely to prevail in the short term

His theory goes something like this.
’Intelligence’ gives power.
Generally that power is not benign - witness man kind discriminating against women and non whites over time because ‘men were more intelligent.
You can then go on to say how we now abuse animals because we are more intelligentl.
He calls this the ‘gorilla effect’. Gorilla’s he argues would say our intelligence is not benign and has caused havoc.

On that basis if AI becomes more intelligent than the human race why on earth wouldn’t it use that intelligence to subjugate/destroy us for its own means?

His belief is, without robust legislation AI will become more intelligent.

Don’t shoot the messenger here - try and find it on BBC Sounds. It was fascinating and just a bit scary.

 

[Laughing]

A brilliant 30 mins of this on R4 this morning (11.00) The guy (Stuart Russell - computer scientist) advises the EU and US government on AI.
His view - the world is at last waking up to a ticking time bomb and he is not optimistic they will to be sensible and regulate will prevail. The economic advantages of AI are likely to prevail in the short term

His theory goes something like this.
’Intelligence’ gives power.
Generally that power is not benign - witness man kind discriminating against women and non whites over time because ‘men were more intelligent.
You can then go on to say how we now abuse animals because we are more intelligentl.
He calls this the ‘gorilla effect’. Gorilla’s he argues would say our intelligence is not benign and has caused havoc.

On that basis if AI becomes more intelligent than the human race why on earth wouldn’t it use that intelligence to subjugate/destroy us for its own means?

His belief is, without robust legislation AI will become more intelligent.

Don’t shoot the messenger here - try and find it on BBC Sounds. It was fascinating and just a bit scary.

I am familiar with Russels thoughts on this subject finny and he raises some interesting talking points, as he generally does.

However, I suspect there was a fair bit missed out of the radio show. They would cover the interesting bits. Most of Russels fears are to do with general AI, and more specifically super general AI. A general AI is considered to be something that can learn anything, in the same way humans do. Feed it data, it learns. If it has the physical attributes to drive a car, it can learn that. Send it to medical school, you get yourself a world class surgeon.

Now if we could develop a general AI, it would become a general super AI very very quickly because of the rapid advancements in computer technology. I don't think moores law holds anymore, but nevertheless things move quickly.

The bit, I am assuming the radio show missed completely, is that we are no where near general AI and we don't know how to get there. It may never happen. We are persisting with the models we do have, which will not solve the problem of general AI.

I didn't listen to the show, but his points about intelligence being destructive is an interesting point. We don't have a lot of examples, humans, and maybe great apes, but you could argue the drive to procreate trumps intelligence for sheer cruelty.

 

[atypical_boro]

"rewrite snow white and the seven dwarves with snow white as an axe murderer"

Once upon a time, in a dark and twisted forest, there lived a wickedly captivating woman named Snow White. She possessed a chilling beauty that lured unsuspecting victims into her web of terror. Instead of being the innocent princess we know, this Snow White had a secret—a penchant for wielding an axe with deadly precision.
Snow White's stepmother, the evil queen, was consumed by envy and despised her stepdaughter's allure. In her twisted mind, she concocted a wicked plan to eliminate Snow White once and for all. But little did the queen know that Snow White had a darkness within her that matched her own.
As the tale unfolds, the queen sent the huntsman to slay Snow White in the heart of the forest. But much to the huntsman's surprise, he found himself face-to-face with a cunning and ruthless Snow White, clutching an axe stained with the blood of her previous victims.
Instead of succumbing to the huntsman's blade, Snow White turned the tables, revealing her true nature. With a sinister smile, she struck down the huntsman, silencing his cries forever.
Snow White knew that her stepmother would never stop hunting her. In order to ensure her survival, she sought refuge deep within the forest, stumbling upon a hidden cottage that belonged to a peculiar group of seven dwarves. Unbeknownst to them, they were about to encounter a force far more dangerous than they could ever imagine.
The dwarves, initially skeptical of Snow White's intentions, soon fell under her enchanting spell. One by one, she manipulated them, using their naivety to her advantage. The cottage walls witnessed the gruesome acts committed by Snow White, as she eliminated anyone who dared to stand in her way.
Yet, despite her sinister nature, the dwarves couldn't help but feel captivated by Snow White's irresistible allure. The lines between fear and fascination blurred, as they became accomplices to her dark deeds.
News of the axe-wielding enchantress spread throughout the kingdom, and even the queen trembled at the thought of facing Snow White's wrath. Consumed by paranoia, the queen disguised herself and ventured into the forest, intending to put an end to Snow White's reign of terror.
However, the queen underestimated the depths of Snow White's malevolence. In a chilling twist, Snow White cunningly lured the queen into a trap, using her beauty to mask the horrors that awaited her. With a swift swing of her axe, Snow White struck the final blow, extinguishing the queen's wicked existence forever.
As the forest fell into an eerie silence, Snow White reveled in her newfound dominance. The dwarves, now fully under her control, served as her loyal subjects. Together, they ruled the enchanted forest with an iron fist, spreading fear and darkness to all who dared to cross their path.
And so, this dark and twisted version of Snow White became a cautionary tale—a chilling reminder that evil can manifest in the most enchanting forms. The legend of Snow White, the axe murderer, continues to haunt the hearts and minds of those who dare to hear her name.

Can it do saucy stuff? Like 'rewrite the story of Macbeth but all the witches are sexy like Catherine Tramell from Basic Instinct'?

Just off the top of my head.....

 

[Centralscrutinizer]

AI has got to be a clear and present danger right now regarding its use to peddle fake news etc regarding politics etc

 

[Littlejimmy]

Can it do saucy stuff? Like 'rewrite the story of Macbeth but all the witches are sexy like Catherine Tramell from Basic Instinct'?

Just off the top of my head.....

Or Romeo and Juliet a la 50 Shades of Grey...

 

[Andy_W]

I think you are taking about different things. @Laughing is correct that ChatGPT doesn't write code. It returns search results depending on what you ask it. Google does that as well but in the form of a website/forum which allows you to read it, analyse it and determine whether the info it has given you is trustworthy, reliable and/or correct. ChatGPT just gives you it as an answer and has no understanding of it. That means you can ask it something very simple and it will return an answer, which might be correct, but without expert knowledge then how can you trust it?

If you are an expert then that simple code can probably be written as quickly as it would take for ChatGPT to give you it and for you to error check. If you aren't an expert you can't trust it. It's more likely to cost you than save you money if you aren't an expert.

So yes, it can give you some basic code but no, it can't write code.

This all might sound odd as I've not dabbled back with this for a couple of months, and would have been better chatting about it back then when I had a rough idea

I'll explain anyway, to give some context, but about a month or so after starting to learn Tkinter GUI basics, along with basic Python on Youtube back in Feb/ March I asked it this (copied directly from my ChatGPT).

"Can you create me a python code, using tkinter, which has 20 labels in column 0, and 20 entry boxes in column 1, and displays them in a GUI window. Add a submit button below this, which then gets the users entries in the boxes, and submits them to a sqlite3 database which it creates."

I could already almost do all of this (very slowly), from what I had learnt, but it wasn't quite integrated correctly and using the database how I wanted (or how I would have needed in the future), and I was a little stuck, but it chucked it out in a minute (and worked), which had took me hours to get to before (obviously still learning).

Then I realised the loop it used was just using label0, label1 etc, as it ran through the loop, so I asked it to use a list of e-mail, name, date etc, so it did, almost identically to how the youtube instructor did etc.

Then I realised it was using a tuple to store data, and I think I wanted a dictionary, so got it to do that instead, which it did, instantly. I don't think I could figure out how to use the dictionary at the time, for this at least, but it did it, and then I asked it to explain and it did, so I got it (since forgot this mind).

Was more than enough to impress me, it blew me away to be honest, and a couple of my mates were impressed how quickly I'd learnt to do it on my own, and get chatgpt to do what I wanted, and how ultimately they both got the same end result and that the code looked fine (for me and chatGPT).

 

[Andy_W]

I had wrote more to explain but it's not posting properly :/ But probably a good thing to stop boring people

 

[Fat_Dragon]

The ChatGPT app is available in UK now for iPhone and iPad users. The Android version is due out at a later date.

When you search for the ChatGPT app you may not find it in the App Store due to copycats trying to get in on the act flooding the app store with versions that make you pay. The official ChatGPT app is free and powered by OpenAI.